Shape Descriptors Based on Polygonal Approximation and Its Computer Simulation
Guangming Liang · Jisuanji fangzhen · 2009
The automatic detection and recognition of cell image is a hotspot in biomedical engineering domain.The shape characters of objects are widely applied in the image identification.Based on the analysis of linear fitting for boundary,an improved image extraction algrithm is proposed,and the curvature vector and length vector for smoothed boundary approximate polygon are established.Then shape descriptors such as number of corners,number of peaks,variance of curvature,maximum chord,vertical chord,chord-ratio,symmetry degree,and parallel degree are defined.The algorithm is applied to process basic shape images and cell images.Computer simulation results show that the extracted features can distinguish basic shape effectively and are invariant in translation.